고급 기능 (Scale to Zero, Lakebase, DABs, AppKit)

# 앱 자동 중지/시작 Job 예시 (Python SDK)
from databricks.sdk import WorkspaceClient

w = WorkspaceClient()

# 퇴근 시 앱 중지 (7PM Job으로 등록)
w.apps.stop("my-dashboard-app")

# 출근 시 앱 시작 (7AM Job으로 등록)
w.apps.start("my-dashboard-app")

# cron 또는 Databricks Job으로 실행
# 퇴근 시 (19:00)
databricks apps stop my-dashboard-app

# 출근 시 (07:00)
databricks apps start my-dashboard-app

# 모든 앱 목록 확인
databricks apps list --output json | jq '.[] | select(.status == "RUNNING") | .name'

# 불필요한 앱 중지
databricks apps stop unused-app-1

{
  "name": "my-scaled-app",
  "description": "Horizontally scaled app",
  "compute_size": "MEDIUM",
  "compute_min_instances": 1,
  "compute_max_instances": 3
}

# Git 기반 배포 (CLI)
databricks apps deploy my-app \
  --git-url https://github.com/myorg/myrepo \
  --git-ref main \
  --source-code-path ./app

command: ['streamlit', 'run', 'app.py']
# 텔레메트리 내보내기 설정 (app.yaml 또는 앱 설정 UI에서)

# pyproject.toml
[project]
name = "my-databricks-app"
version = "0.1.0"
requires-python = ">=3.10"
dependencies = [
    "streamlit>=1.32.0",
    "databricks-sdk>=0.20.0",
    "pandas>=2.2.0",
]

# uv.lock 생성
uv lock

command: ['python', 'app.py']
env:
  - name: DB_CONNECTION
    valueFrom: lakebase_db
resources:
  - name: lakebase_db
    type: postgres

# requirements.txt
databricks-sdk
psycopg2-binary    # PostgreSQL 드라이버

import os
import psycopg2

# Lakebase 연결 정보 가져오기
db_connection = os.getenv("DB_CONNECTION")

# PostgreSQL 호환 연결
conn = psycopg2.connect(db_connection)

# CRUD 작업
with conn.cursor() as cur:
    # 테이블 생성
    cur.execute("""
        CREATE TABLE IF NOT EXISTS app_sessions (
            id SERIAL PRIMARY KEY,
            user_id VARCHAR(255),
            data JSONB,
            created_at TIMESTAMP DEFAULT NOW()
        )
    """)

    # 데이터 삽입
    cur.execute(
        "INSERT INTO app_sessions (user_id, data) VALUES (%s, %s)",
        ("user123", '{"page": "dashboard", "action": "view"}')
    )
    conn.commit()

    # 데이터 조회
    cur.execute("SELECT * FROM app_sessions WHERE user_id = %s", ("user123",))
    rows = cur.fetchall()

conn.close()

import os
import streamlit as st
import psycopg2

@st.cache_resource
def get_db_connection():
    """Lakebase 연결을 캐싱하여 재사용합니다."""
    return psycopg2.connect(os.getenv("DB_CONNECTION"))

conn = get_db_connection()

st.title("피드백 수집 앱")

# 피드백 입력 폼
with st.form("feedback_form"):
    name = st.text_input("이름")
    feedback = st.text_area("피드백")
    rating = st.slider("만족도", 1, 5, 3)
    submitted = st.form_submit_button("제출")

    if submitted:
        with conn.cursor() as cur:
            cur.execute(
                "INSERT INTO feedback (name, feedback, rating) VALUES (%s, %s, %s)",
                (name, feedback, rating)
            )
            conn.commit()
        st.success("피드백이 저장되었습니다!")

# 피드백 목록 조회
with conn.cursor() as cur:
    cur.execute("SELECT name, feedback, rating, created_at FROM feedback ORDER BY created_at DESC LIMIT 20")
    rows = cur.fetchall()

if rows:
    import pandas as pd
    df = pd.DataFrame(rows, columns=["이름", "피드백", "만족도", "작성일"])
    st.dataframe(df)

my-app-bundle/
├── databricks.yml          # Bundle 설정 파일
├── src/
│   ├── app.py              # 앱 코드
│   ├── app.yaml            # 앱 런타임 설정
│   └── requirements.txt    # 의존성
└── resources/
    └── my_app.yml           # 앱 리소스 정의

bundle:
  name: my-dashboard-app

workspace:
  host: https://<workspace-url>

targets:
  dev:
    mode: development
    default: true
    workspace:
      host: https://<dev-workspace-url>

  prod:
    mode: production
    workspace:
      host: https://<prod-workspace-url>

include:
  - resources/*.yml

resources:
  apps:
    my_dashboard_app:
      name: "my-dashboard-app"
      description: "매출 분석 대시보드"
      source_code_path: ../src
      config:
        command:
          - streamlit
          - run
          - app.py
        env:
          - name: DATABRICKS_WAREHOUSE_ID
            value: ${var.warehouse_id}
      resources:
        - name: sql_warehouse
          sql_warehouse:
            id: ${var.warehouse_id}
            permission: CAN_USE
      permissions:
        - level: CAN_MANAGE
          user_name: admin@company.com
        - level: CAN_USE
          group_name: data-team

# Bundle 유효성 검사
databricks bundle validate -t dev

# 개발 환경에 배포
databricks bundle deploy -t dev

# 프로덕션 환경에 배포
databricks bundle deploy -t prod

# 배포 상태 확인
databricks bundle summary -t dev